Clinical utility of tumor markers in early diagnosis of ...

Clinical utility of tumor markers in early

diagnosis of hepatocellular carcinoma

Tae Seop Lim

Department of Medicine

The Graduate School, Yonsei University



Tae Seop Lim





Directed by Professor Do Young Kim

The Master's Thesis submitted to the Department of

Medicine, the Graduate School of Yonsei University

in partial fulfillment of the requirements for the

degree of Master of Medical Science

Tae Seop Lim

December 2013

This certifies that the Master's Thesis of

Tae Seop Lim is approved.

------------------------------------ Thesis Supervisor : Do Young Kim

------------------------------------ Thesis Committee Member #1 : Kwang-Hyub Han

------------------------------------ Thesis Committee Member #2 : Hyon-Suk Kim

The Graduate School

Yonsei University

December 2013

ACKNOWLEDGEMENTS

I would like to express my deep gratitude to my supervisor,

Professor Do Young Kim, who gave me his exemplary guidance,

monitoring, and constant encouragement through the course of

this thesis.

I would also like to convey thanks to Professor, Kwang-Hyub

Han, and Professor Hyon-Suk Kim for professional guidance and

valuable support in completing this work.

Lastly, I wish to thank my parents and colleagues for their

support and encouragement throughout my study.

.

<TABLE OF CONTENTS>

ABSTRACT ······························································································ 1

I. INTRODUCTION ·················································································· 3

II. MATERIALS AND METHODS

1. Patients ···························································································· 5

2. Measurements of tumor markers ·················································· 6

3. HCC staging system ········································································ 6

4. Statistical analyses ·········································································· 6

III. RESULTS

1. Baseline characteristics of patients ··················································· 7

2. Diagnostic accuracy of tumor markers to detect overall HCC ··················· 7

3. Diagnostic accuracy of tumor markers to distinguish early HCC············· 13

4. Diagnostic utility of PIVKA-II and AFP-L3 in patients with AFP <20 ng/ml

············································································································ 16

5. HCC diagnosis probability in patients with AFP ≥ 20ng/ml ················ 18

6. Correlation between tumor-related variables and serum level of tumor markers

········································································································· 20

IV. DISCUSSION ·················································································· 22

V. CONCLUSION ················································································· 25

REFERENCES ······················································································ 25

ABSTRACT (IN KOREAN) ································································ 33

LIST OF FIGURES

Figure 1. Comparison of serum AFP, PIVKA-II, and AFP-L3 values

in patients with HCC and LC ················································· 10

Figure 2. ROC curves of AFP, PIVKA-II, and AFP-L3 for

distinguishing HCC from LC ··················································· 11

Figure 3. ROC curves of AFP, PIVKA-II, and AFP-L3 for

distinguishing early HCC (single tumor less than 3 cm in size) from

LC ····························································································· 14

Figure 4. ROC curves of PIVKA-II and AFP-L3 for distinguishing

HCC from LC in patients with AFP <20 ng/ml ······················ 17

Figure 5. The graph of HCC diagnosis probability as AFP-L3 level in

patients with AFP ≥ 20ng/ml ················································ 19

LIST OF TABLES

Table 1. Baseline characteristics of the study population. ········· 8

Table 2. Sensitivity, specificity, PPV, and NPV for different cut-off

values of tumor markers in distinguishing overall HCC from 12

Table 3. Sensitivity, specificity, PPV, and NPV for different cut-off

values of tumor markers in distinguishing early HCC (single tumor

less than 3 cm in size) from LC ··············································· 15

Table 4. Sensitivity, specificity, PPV, and NPV for PIVKA-II and

AFP-L3 in patients with AFP <20ng/ml ·································· 18

Table 5. Logistic regression analysis for HCC diagnosis probability

in patients with AFP ≥ 20ng/ml ············································ 20

Table 6. Relationship between tumor-related variables and tumor

marker levels ············································································ 21

1

<ABSTRACT>

Early diagnosis of hepatocellular carcinoma (HCC) is very important for a

favorable prognosis. Some serologic tests including alpha-fetoprotein (AFP),

protein induced by vitamin K absence-II (PIVKA-II), and lens culinaris

agglutinin-reactive fraction of AFP (AFP-L3) have been studied as diagnostic

markers of HCC; however, there is no consensus on which tumor markers are

the most effective in detecting early HCC. In this study, we investigate the

clinical utility of tumor markers in the early diagnosis of HCC. A total of 425

patients with liver cirrhosis (LC) (n =196) or HCC (n = 229) were studied

from January 2012 to February 2013. Patients with LC had a mean age of 55.8

years and 58.7% were male, whereas the mean age of patients with HCC was

60.0 years and 76.0% were male. We analyzed the expression of tumor

markers AFP, PIVKA-II, and AFP-L3 in these patients. All tumor markers

were significantly elevated in HCC patients compared with LC patients (p

<0.001). The area under the receiver operating characteristic curves (AUROC)

of AFP, PIVKA-II, and AFP-L3 for distinguishing HCC from LC was 0.679

(95% confidence interval [CI], 0.626-0.732, p <0.001), 0.812 (95% CI, 0.770-

0.854, p <0.001), and 0.690 (95% CI, 0.638-0.742, p <0.001), respectively.

Moreover, PIVKA-II (AUROC = 0.705, 95% CI, 0.621-0.789, p <0.001) was

superior to AFP (AUROC = 0.623, 95% CI, 0.527-0.719, p = 0.019) and AFP-

2

L3 (AUROC = 0.561, 95% CI, 0.453-0.668, p = 0.245) for diagnosis of early

HCC, defined as a single tumor less than 3 cm in size. The low sensitivity

(25.6%) of PIVKA-II (cut-off 40 mAU/ml) can be overcome by combining it

with AFP (48.7%). Furthermore, with combined AFP (cut-off 20 ng/ml),

PIVKA-II (cut-off 40 mAU/ml), and AFP-L3 (cut-off 10%), the sensitivity

was enhanced to 56.4%. In patients with AFP <20 ng/ml, the AUROC for

PIVKA-II (0.743, 95% CI, 0.678-0.807; p = <0.001) was superior to that of

AFP-L3 (0.576, 95% CI, 0.500-0.653; p = 0.052). AFP-L3 was able to

differentiate HCC patients from AFP-false negative patients in the logistic

regression analysis (odds ratio 1.076, 95% CI, 1.037-1.116, p = < 0.001). All

tumor marker levels including AFP, PIVKA-II, and AFP-L3 correlated with

the size and stage of HCC with statistical significance. In conclusion,

combined AFP and PIVKA-II can used for good screening tool of early HCC.

Furthermore, AFP-L3 may have an additional role to differentiate between true

HCC in AFP false-positive patients.

----------------------------------------------------------------------------------------

Key words: Hepatocellular carcinoma, AFP, PIVKA-II, AFP-L3, Tumor

marker

3

Clinical utility of tumor markers in early diagnosis of hepatocellular

carcinoma

Tae Seop Lim



(Directed by Professor Do Young Kim)

I. INTRODUCTION

Hepatocellular carcinoma (HCC) is the sixth most common cancer in the world and

the third most common cause of cancer-related death.1 Because the prognosis of HCC

is usually determined by stage at the time of diagnosis, early diagnosis of HCC is very

important.

Serum alpha-fetoprotein (AFP) is the most representative tumor marker for HCC,

and is used in the diagnosis and surveillance of HCC in combination with abdominal

ultrasonography. A cutoff value of 20 ng/ml AFP is most commonly used.2

Furthermore, patients with serum AFP greater than 400 ng/ml were reported to have

greater tumor size, portal vein thrombosis, diffuse or massive types, and a lower

survival rate.3,4

However, because AFP level can also be increased in chronic hepatitis

or liver cirrhosis without HCC5,6

there is a need for markers with higher specificity.

Protein induced by vitamin K absence-II (PIVKA-II) is produced as the result of a

defect in posttranslational carboxylation of the prothrombin precursor in cancer cells

and shows abnormally increased expression in HCC patients. PIVKA-II has therefore

been considered as an additional marker for the diagnosis of HCC, and some previous

4

studies have reported that PIVKA-II is more accurate than AFP in the diagnosis of

HCC.7-10

Lens culinaris agglutinin-reactive fraction of AFP (AFP-L3) is an isoform of AFP

that reflects changes in the carbohydrate chain and is also a specific marker for

HCC.11

Furthermore, AFP-L3 has been shown to be useful in early diagnosis,11,12

prognosis after treatment,13,14

or prediction of malignant potential in HCC.15,16

Previous studies suggested 10% as the cutoff value of AFP-L3.2 The analytical

sensitivity of AFP-L3 is affected by total AFP level because AFP-L3 is described as a

percentage of total AFP level. Until recently the percentage of AFP-L3 has been

analyzed by liquid-phase binding assay (LBA); however, this method is affected by

the total AFP level and its clinical utility is low in patients with low total AFP.17,18

A

highly sensitive AFP-L3 assay using micro-total analysis systems (μTAS) was

suggested to overcome the limited value of the conventional method. This new

method can measure AFP-L3 accurately in patients with very low AFP level,19

and a

recent study indicated that AFP-L3 measured by this highly sensitive technique was a

more useful marker for diagnosis and predicting prognosis in HCC patients.20

Although AFP, PIVKA-II, and AFP-L3 are all recognized tumor markers for HCC,

there is no consensus on which tumor marker is the most useful indicator of early

diagnosis. Therefore, in this study we determined which tumor marker is the most

useful for early diagnosis in HCC.

5

II. MATERIALS AND METHODS

1. Patients

Between January 2012 and February 2013, 229 consecutive patients who were

diagnosed with HCC for the first time at Severance Hospital, Yonsei University

College of Medicine, Seoul, Korea were enrolled in this study. HCC was diagnosed

histologically or radiologically according to the guidelines of the American

Association for the Study of Liver Disease (AASLD) or the European Association for

Study of the Liver (EASL). The clinical diagnosis was made based on typical

radiologic findings in dynamic computed tomography (CT), dynamic magnetic

resonance imaging (MRI), or hepatic angiography in combination with increased AFP

greater than 200 ng/ml in patients with chronic hepatitis or liver cirrhosis. If the AFP

level was less than 200 ng/ml, at least two imaging findings should be consistent with

HCC.21-23

The exclusion criteria were as follows: age younger than 18 years;

previously diagnosed HCC; previous history of liver transplantation or liver resection;

any cancers other than HCC. Control samples were obtained from 196 consecutive

cirrhotic patients without HCC between January 2012 and February 2013. Liver

cirrhosis was defined by histology, clinical, biochemical, or imaging findings. This

study protocol was approved by the institutional ethics review board and was in

compliance with the Declaration of Helsinki.

6

2. Measurements of tumor markers

AFP, PIVKA-II, and AFP-L3 were measured in serum samples obtained from LC and

HCC patients. For HCC patients, serum samples were collected at the time of HCC

diagnosis before treatment. Measurements of AFP, PIVKA-II, and AFP-L3 were

performed using the μTAS assay (Wako Pure Chemical Industries, Ltd, Osaka,

Japan).19

3. HCC staging system

The American Joint Committee on Cancer (AJCC) TNM (tumor-node-metastasis) and

Barcelona Clinic Liver Cancer (BCLC) systems were used for HCC staging.

4. Statistical analyses

Continuous variables were compared with t-tests or Mann-Whitney U tests, and

categorical variables were compared using the chi-squared or Fisher’s exact tests.

Receiver operating characteristic (ROC) curves were constructed and the areas under

the ROC curves (AUROC) were calculated. The sensitivity, specificity, positive

predictive value (PPV), and negative predictive value (NPV) were calculated. The

relationship between AFP-L3 and HCC diagnosis probability in patients with an AFP

level greater than 20ng/ml were analyzed with a logistic regression model. A

7

probability (p) value of 0.05 was chosen for statistical significance. Statistical

analyses were performed using SPSS version 18.0 (SPSS Inc., Chicago, IL, USA).

III. Results

1. Baseline characteristics of patients

A total of 425 patients with LC (n = 196) or HCC (n = 229) were enrolled. The

patient’s baseline characteristics are described in Table 1. Patients with LC had a

mean age of 55.8 years and 58.7% were male. Patients with HCC had a mean age of

60.0 years and 76.0% were male. The 229 patients with HCC included 75 (32.3%) in

stage I, 53 (23.1%) in stage II, 55 (24.0%) in stage III, and 45 (19.7%) in stage IV.

Serum levels of AFP, PIVKA-II, and AFP-L3 were all significantly elevated in HCC

patients compared with LC patients. Comparison of tumor markers in patients with

LC and HCC is shown in Figure 1.

2. Diagnostic accuracy of tumor markers to detect overall HCC

The ROC curves of tumor markers for distinguishing HCC from LC are shown in

Figure 2. AUROC for AFP, PIVKA-II, and AFP-L3 was 0.679 (95% CI, 0.626-0.732,

p <0.001), 0.812 (95% CI, 0.770-0.854, p <0.001), and 0.690 (95% CI, 0.638-0.742, p

<0.001), respectively. The sensitivity, specificity, NNP, and PPV for different cut-off

8

values are presented in Table 2. The three tumor markers combined resulted in an

enhanced sensitivity of 80.8%.

Table 1. Baseline characteristics of the study population.

Variables Patients with LC (n = 196) Patients with HCC (n = 229) p

Male gender (%) 115 (58.7%) 174 (76.0%)

Age (years) 55.8 ± 10.7 60.0 ± 10.9 <0.001

Etiology

HBV/HCV/ 122 (62.2%) /33 (16.8%) 174 (76.0%) /23 (10.0%)

Alcohol/Others, n (%) /23 (11.7%) /18 (9.2%) /12 (5.2%) /21 (9.2%)

Child-Pugh class: 171 (87.2%) /17 (8.7%) 184 (80.3%) /41 (17.9%)

A/B/C, n (%) /8 (4.0%) /5 (2.2%)

Hb (g/dl) 13.6 (12.3-15.0) 13.3 (11.8-14.6) 0.176

Platelet count (× 103/mm3) 116.5 (80.5-149.7) 159.5 (100.8-205.3) <0.001

AST (IU/L) 37.5 (28.0-56.0) 43.5 (29.0-81.0) 0.005

ALT (IU/L) 31.0 (20.0-44.8) 32.0 (20.8-54.3) 0.134

Albumin (g/dl) 4.2 (3.7-4.5) 3.8 (3.2-4.2) <0.001

Total bilirubin (mg/dl) 0.9 (0.7-1.2) 0.9 (0.6-1.3) 0.327

PT INR 1.0 (1.0-1.1) 1.0 (1.0-1.1) 0.047

AFP (ng/ml) 4.5 (2.4-16.2) 22.0 (5.1-628.3) <0.001

PIVKA-II (mAU/ml) 19.0 (14.3-27.0) 85.5 (24.0-2000.0) <0.001

AFP-L3 (%) 2.1 (0.0-6.9) 8.7 (1.2-32.9) <0.001

Vessel invasion (%) NA 62 (27.1%)

Portal vein thrombosis (%) NA 34 (14.8%)

Distant metastasis (%) NA 20 (8.7%)

9

Tumor number, ≥2 (%) NA 90 (39.3%)

Tumor size, ≥3 cm (%) NA 147 (64.2%)

TNM staging: NA 74 (32.3%) /53 (23.1%)

I/II/III/IV, n (%)

/55 (24.0%) /45 (19.7%)

BCLC staging: NA 88 (38.4%) /43 (18.8%)

A/B/C/D, n (%)

/92 (40.2%) /7 (3.1%)

Data are expressed as the number (percentage), mean±SD, and median (interquartile

range).

HBV, Hepatitis B virus; HCV, Hepatitis C virus; AST, Aspartate aminotransferase;

ALT, Alanine aminotransferase; PT INR, Prothrombin time international normalized

ratio; AFP, Alpha-fetoprotein; PIVKA-II, Protein induced by vitamin K absence-II;

AFP-L3, Lens culinaris agglutinin-reactive fraction of alpha-fetoprotein; TNM,

Tumor-node-metastasis; BCLC, Barcelona Clinic Liver Cancer

10

(a) (b)

LC HCC0.1

1

10

100

1000

10000

100000

1000000

P = <0.001

AF

P (

ng

/ml)

LC HCC1

10

100

1000

10000

100000

P = <0.001

PIV

KA

-II (m

AU

/ml)

(c)

LC HCC0

20

40

60

80

100

P = <0.001

AF

P-L

3 (

%)

Figure 1. Comparison of serum AFP, PIVKA-II, and AFP-L3 values in patients with

HCC and LC. The values of AFP (a), PIVKA-II (b), and AFP-L3 (c) are shown as

rectangles, in which the line represents the median.

LC, Liver cirrhosis; HCC, Hepatocellular carcinoma; AFP, Alpha-fetoprotein;

PIVKA-II, Protein induced by vitamin K absence-II; AFP-L3, Lens culinaris

agglutinin-reactive fraction of alpha-fetoprotein

11

Figure 2. ROC curves of AFP, PIVKA-II, and AFP-L3 for distinguishing HCC from

LC. AUROC was 0.679 (95% CI, 0.626-0.732, p <0.001) for AFP, 0.812 (95% CI,

0.770-0.854, p <0.001) for PIVKA-II, and 0.690 (95% CI, 0.638-0.742, p <0.001) for

AFP-L3.

ROC, Receiver operating characteristic; LC, Liver cirrhosis; HCC, Hepatocellular

carcinoma; AUROC, The area under the receiver operating characteristic curves;

AFP, Alpha-fetoprotein; PIVKA-II, Protein induced by vitamin K absence-II; AFP-

L3, Lens culinaris agglutinin-reactive fraction of alpha-fetoprotein

12

Table 2. Sensitivity, specificity, PPV, and NPV for different cut-off values of tumor

markers in distinguishing overall HCC from LC

Variables Cut-off value Sensitivity Specificity NPV PPV

AFP 20 ng/ml 52.6% 78.6% 58.6% 74.2%

200 ng/ml 30.4% 98.0% 54.5% 94.6%

PIVKA-II 40 mAU/ml 59.6% 91.9% 66.0% 89.5%

100 mAU/ml 49.1% 96.0% 61.6% 93.3%

AFP-L3 5% 60.8% 65.3% 58.7% 67.3%

10% 46.1% 89.7% 58.5% 84.1%

AFP +PIVKA-II 20ng/ml for AFP or

40mAU/ml for PIVKA-II

76.4% 71.9% 72.3% 76.0%

PIVKA-II+AFP-L3 40mAU/ml for PIVKA-II or

10% for AFP-L3

71.3% 85.2% 71.7% 85.0%

AFP + PIVKA –II

+ AFP-L3

20ng/ml for AFP,

40mAU/ml for PIVKA-II,

or 10% for AFP-L3

80.8% 67.3% 75.0% 74.3%

PPV, Positive predictive value; NPV, Negative predictive value; AFP, Alpha-

fetoprotein; PIVKA-II, Protein induced by vitamin K absence-II; AFP-L3, Lens

culinaris agglutinin-reactive fraction of alpha-fetoprotein

13

3. Diagnostic accuracy of tumor markers to distinguish early HCC

To evaluate the diagnostic accuracy of tumor markers in distinguishing early HCC

from LC, we analyzed ROC curves as shown in Figure 3. Early HCC was defined as a

single tumor less than 3 cm in size. The number of patients with the early HCC was

39. The AUROC that diagnosed patients with early HCC was 0.623 (95% CI, 0.527-

0.719, p = 0.019) for AFP, 0.705 (95% CI, 0.621-0.789, p = 0.001) for PIVKA-II, and

0.561 (95% CI, 0.453-0.668, p = 0.245) for AFP-L3. The Sensitivity, specificity, PPV,

and NPV for different cut-off values of tumor markers in distinguishing early HCC

from LC is described in table 3. The low sensitivity (25.6%) of PIVKA-II (cut-off 40

mAU/ml) can be overcome by combining it with AFP (48.7%). Furthermore, with

combination of three tumor markers, the sensitivity was enhanced to 56.4%.

14

Figure 3. ROC curves of AFP, PIVKA-II, and AFP-L3 for distinguishing early HCC

(single tumor less than 3 cm in size) from LC. AUROC was 0.623 (95% CI, 0.527-

0.719, p = 0.019) for AFP, 0.705 (95% CI, 0.621-0.789, p <0.001) for PIVKA-II, and

0.561 (95% CI, 0.453-0.668, p = 0.245) for AFP-L3.


carcinoma; AFP, Alpha-fetoprotein; PIVKA-II, Protein induced by vitamin K

absence-II; AFP-L3, Lens culinaris agglutinin-reactive fraction of AFP; AUROC,

The area under the receiver operating characteristic curves

15

Table 3. Sensitivity, specificity, PPV, and NPV for different cut-off values of tumor

markers in distinguishing early HCC (single tumor less than 3 cm in size) from LC


AFP 20 ng/ml 41.0% 78.6% 87.0% 27.6%

200 ng/ml 12.8% 98.0% 85.0% 55.6%

PIVKA-II 40 mAU/ml 25.6% 91.8% 86.1% 38.5%

100 mAU/ml 12.8% 95.9% 84.7% 38.4%

AFP-L3 5% 43.6% 65.3% 85.3% 20.0%

10% 25.6% 85.8% 85.8% 33.3%

AFP +PIVKA-II 20ng/ml for AFP or


48.7% 71.9% 87.6% 25.7%

PIVKA-II+AFP-

L3


or 10% for AFP-L3

41.0% 85.2% 87.9% 35.6%

AFP + PIVKA–II

+ AFP-L3

20ng/ml for AFP,

40mAU/ml for PIVKA-II,

or 10% for AFP-L3

56.4% 67.3% 88.6% 25.9%



culinaris agglutinin-reactive fraction of AFP

16

4. Diagnostic utility of PIVKA-II and AFP-L3 in patients with AFP <20 ng/ml

The ROC curves of PIVKA-II and AFP-L3 in patients with AFP <20 ng/ml is given

in Figure 4. The number of LC and HCC patients with AFP <20 ng/ml was 197 and

66, respectively. The diagnostic accuracy of PIVKA-II was superior to that of AFP-

L3 in patients with AFP <20 ng/ml. The AUROC of PIVKA and AFP-L3 was 0.743

(95% CI, 0.678-0.807, p <0.001) and 0.576 (95% CI, 0.500-0.653, p = 0.052),

respectively. The sensitivity, specificity, PPV, and NPV for PIVKA-II and AFP-L3 in

patients with an AFP <20 ng/ml is presented in Table 4. The sensitivity of PIVKA-II,

with a cut-off value of 40 mAU/ml, was 48.6%, and PIVKA-II and AFP-L3 combined

showed an enhanced sensitivity of up to 57.8%.

17

Figure 4. ROC curves of PIVKA-II and AFP-L3 for distinguishing HCC from LC in

patients with AFP <20 ng/ml. AUROC was 0.743 (95% CI, 0.678-0.807, p <0.001)

for PIVKA-II and 0.576 (95% CI, 0.500-0.653, p = 0.052) for AFP-L3.


carcinoma; AFP, Alpha-fetoprotein; AUROC, The area under the receiver

operating characteristic curves; PIVKA-II, Protein induced by vitamin K absence-

II; AFP-L3, Lens culinaris agglutinin-reactive fraction of alpha-fetoprotein

18

Table 4. Sensitivity, specificity, PPV, and NPV for PIVKA-II and AFP-L3 in patients

with AFP < 20ng/ml


PIVKA-II 40 mAU/ml 48.6% 91.6% 71.6% 80.3%

100 mAU/ml 37.6% 96.1% 68.5% 87.2%

AFP-L3 5% 42.2% 70.1% 63.2% 50.0%

10% 24.7% 91.5% 63.0% 67.5%

PIVKA-II+AFP-L3 40mAU/ml for PIVKA-

II or 10% for AFP-L3

57.8% 85.7% 74.2% 74.1%



culinaris agglutinin-reactive fraction of alpha-fetoprotein

5. HCC diagnosis probability in patients with AFP ≥ 20ng/ml

We present HCC diagnostic probability in patients with AFP ≥ 20ng/ml in Figure 5.

HCC diagnostic probability of AFP-L3 was calculated with univariate logistic

regression analysis (Table 5). The number of LC and HCC patients with AFP ≥

20ng/ml was 42 and 121, respectively. Among the LC patients with a false-positive

AFP (AFP ≥ 20ng/ml), 35 (83.3%) LC patients had an AFP-L3 less than 10%. HCC

19

diagnosis probability and AFP-L3 level were correlated with statistical significance (p

< 0.001).

Figure 5. The graph of HCC diagnosis probability as AFP-L3 level in patients with

AFP ≥ 20ng/ml

HCC, Hepatocellular carcinoma; AFP, Alpha-fetoprotein; AFP-L3, Lens culinaris

agglutinin-reactive fraction of alpha-fetoprotein

20

Table 5. Logistic regression analysis for HCC diagnosis probability in patients with

AFP ≥ 20ng/ml

OR 95% CI p

AFP-L3 (%) 1.076 1.037 - 1.116 <0.001

HCC, Hepatocellular carcinoma; AFP-L3, Lens culinaris agglutinin-reactive fraction

of alpha-fetoprotein; OR, Odds ratio; CI, Confidence interval

6. Correlation between tumor-related variables and serum level of tumor markers

We investigated the correlation between tumor markers and HCC staging and found

that levels of AFP, PIVKA, and AFP-L3 were significantly elevated in HCC larger

than 3 cm, in the presence of vascular invasion or distant metastasis, and in disease

with AJCC stage III and IV (Table 6). All tumor markers correlated with tumor size

and staging with statistical significance.

21

Table 6. Relationship between tumor-related variables and tumor marker levels

Variables AFP (ng/ml) PIVKA-II (mAU/ml) AFP-L3 (%)

Tumor size (cm)

≤3 15.0 (5.1-77.8) 27.0 (20.0-57.0) 4.9 (0.0-12.1)

>3 31.1 (5.1-1663.0) 775.0 (44.0-2528.0) 13.7 (2.4-48.4)

p value 0.014 <0.001 <0.001

Vascular invasion

No 20.6 (5.1-170.3) 42.0 (22.0-401.3) 6.2 (0.0-22.6)

Yes 115.2 (5.0-4784.5) 2000.0 (149.0-2801.5) 27.0 (5.7-55.1)

p value 0.013 <0.001 <0.001

Distant metastasis

No 19.7 (4.8-257.7) 58.5 (23.0-1485.8) 7.5 (0.0-28.8)

Yes 4338.2 (251.1-48473.6) 20000.0 (1844.0-9441.3) 33.3 (10.4-71.5)

p value <0.001 <0.001 0.001

TNM stage

I+II 15.2 (5.1-189.9) 32.0 (20.3-134.5) 6.0 (0.0-21.0)

III+IV 44.3 (5.2-1772.5) 1792.0 (126.0-3343.0) 16.4 (2.7-56.9)

p value 0.033 <0.001 <0.001

AFP, Alpha-fetoprotein; PIVKA-II, Protein induced by vitamin K absence-II; AFP-

L3, Lens culinaris agglutinin-reactive fraction of alpha-fetoprotein; TNM, Tumor-

node-metastasis

22

IV. DISCUSSION

AFP, PIVKA-II, and AFP-L3 are commonly used as tumor markers for HCC. Of

these, AFP is the most widely used marker for monitoring HCC development, and

AFP assessment and liver ultrasonography are the most commonly used tools for

HCC surveillance. However, elevated AFP is not known to be reliable in patients with

early HCC or a small tumor,24-28

and the performance of ultrasonography depends on

several factors such as the examiner’s experience, the technology used, the patient’s

body habitus, the presence of cirrhosis, and tumor size.29

Ultrasonography has a

particularly low sensitivity for detecting tumor nodules in a cirrhotic liver.30-32

Despite

these limitations, there are no reliable prospective data on other tumor markers such

as PIVKA-II and AFP-L3, therefore AFP and ultrasonography are still being

commonly used for HCC surveillance.29

PIVKA-II was first described as a tumor marker of HCC by Liebman et al. in

1984.33

PIVKA-II is a more specific marker than total AFP in the diagnosis for HCC

because other liver diseases rarely give rise to elevated PIVKA-II.10,34

Although an

American study suggested that PIVKA-II was significantly better than AFP or AFP-

L3 in differentiating HCC from cirrhosis for total HCC and small HCC,35

a Japanese

study demonstrated that PIVKA-II has limited value in detecting small HCC. In the

latter study, the efficacy of PIVKA-II was lower than that of AFP in the diagnosis of

HCC smaller than 3 cm, whereas the opposite result was obtained for tumors larger

than 5 cm.36

In the present study, PVIKA-II was superior to AFP and AFP-L3 for not

only detecting overall HCC, but also for detecting early stage HCC. This finding is

23

consistent with a recent Korean study, which reported that PIVKA-II was a more

useful marker than AFP for differentiating HCC from liver cirrhosis, especially in

cases with small HCC.37

Although PIVKA-II has been reported to be more sensitive

than AFP in HCC diagnosis in North America and East Asian countries,38-40

European

studies have shown different results. These discrepancies may be related to etiologic

factors in addition to racial factors.29

Furthermore, the control group in our study was

limited to patients with liver cirrhosis. Two studies suggesting that PIVKA-II was

better than AFP included a control group limited to patients with liver cirrhosis,35,37

whereas one study showing that PIVKA-II was not useful for detecting early HCC

included patients with chronic liver disease, with or without LC.36

Although the

mechanism is currently unknown, PIVKA-II may be the most specific marker for

detecting early HCC especially with respect to liver cirrhosis, and subgroup analysis

including chronic hepatitis without liver cirrhosis will be needed.

Ultrasonography has a limited role in the detection of early HCC especially in

cirrhotic liver;30-32

therefore, increasing the sensitivity of tumor markers is very

important to diagnose early HCC. Although PIVKA-II showed the best diagnostic

accuracy for the detection of early HCC in our study, its sensitivity (cut-off

40mAU/ml) was only 25.6%. This weak point can be overcome when it is combined

with AFP. The sensitivity of combined PIVKA-II (cut-off 40mAU/ml) and AFP (cut-

off 20ng/ml) was enhanced to 48.7%, and the sensitivity of the three tumor markers

combined was enhanced to 56.4%. In this study, PPV to detect early HCC was only

24

25.9%, despite combining the three tumor markers. We think this is due to our small

sample size of early HCC patients (n = 39).

In the present study, PIVKA-II was also superior to AFP-L3 for the diagnosis of

HCC in patients with AFP <20 ng/ml. The AUROC of AFP-L3 was 0.576 without

statistical significance. This finding is not consistent with recent Japanese studies that

reported the usefulness of AFP-L3 in patients with serum AFP less than 20 ng/ml.20,41

In Japan, hepatitis C virus is the most common etiology, while hepatitis B virus is the

most common cause of HCC in Korea. Moreover, these Japanese studies defined the

control group as chronic liver disease regardless of liver cirrhosis. We think that

discrepancies between our study and recent Japanese studies are due to different

etiologies and control groups.

AFP is the most commonly used tumor marker in high-risk patients of HCC, but it

has limited value due to its low specificity.5,6

In comparison with AFP, AFP-L3 has

been known as a very specific marker for HCC.42

In this study, the proportion of LC

patients with AFP-false positive (cut-off 20ng/ml) was 21.4% (42/196). In patients

with AFP ≥ 20ng/ml, the probability of HCC diagnosis was significantly increased as

AFP-L3 levels. This finding can suggest that AFP-L3 can differentiate between true

HCC in AFP false-positive patients.

In this study, serum levels of the three tumor markers all showed a correlation with

tumor size and staging with statistical significance. Although serum AFP is markedly

elevated in patients with distant metastasis, the median value of AFP was less than

200 ng/ml not only in early-stage disease, but also in patients with tumors larger than

25

3 cm, vascular invasion, and advanced TNM stage. This is interesting because clinical

diagnosis was defined as AFP greater than 200 ng/ml. Our findings suggest that other

tumor markers such as PIVKA-II or AFP-L3 might be needed to evaluate treatment

response in both early and advanced stage HCC.

V. CONCLUSION

Combined AFP and PIVKA-II can used for good screening tool of early HCC.

Furthermore, AFP-L3 may have an additional role to differentiate between true HCC

in AFP false-positive patients.

REFERENCE

1. Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002. CA: a

cancer journal for clinicians 2005;55:74-108.

2. Durazo FA, Blatt LM, Corey WG, Lin JH, Han S, Saab S, et al.

Des‐γ‐carboxyprothrombin, α‐fetoprotein and AFP‐L3 in patients with

chronic hepatitis, cirrhosis and hepatocellular carcinoma. Journal of

gastroenterology and hepatology 2008;23:1541-8.

3. Tangkijvanich P, Anukulkarnkusol N, Suwangool P, Lertmaharit S,

Hanvivatvong O, Kullavanijaya P, et al. Clinical characteristics and prognosis

of hepatocellular carcinoma: analysis based on serum alpha-fetoprotein levels.

Journal of clinical gastroenterology 2000;31:302-8.

26

4. Fujioka M, Nakashima Y, Nakashima O, Kojiro M. Immunohistologic study

on the expressions of α‐fetoprotein and protein induced by vitamin K absence

or antagonist II in surgically resected small hepatocellular carcinoma.

Hepatology 2003;34:1128-34.

5. Bloomer JR. Serum alpha-fetoprotein in nonneoplastic liver diseases.

Digestive Diseases and Sciences 1980;25:241-2.

6. Alpert E, Feller E. Alpha-fetoprotein (AFP) in benign liver disease. Evidence

that normal liver regeneration does not induce AFP synthesis.

Gastroenterology 1978;74:856.

7. Mita Y, Aoyagi Y, Yanagi M, Suda T, Suzuki Y, Asakura H. The usefulness

of determining des‐γ‐carboxy prothrombin by sensitive enzyme immunoassay

in the early diagnosis of patients with hepatocellular carcinoma. Cancer

2000;82:1643-8.

8. Nakagawa T, Seki T, Shiro T, Wakabayashi M, Imamura M, Itoh T, et al.

Clinicopathologic significance of protein induced vitamin K absence or

antagonist II and alpha-fetoprotein in hepatocellular carcinoma. International

journal of oncology 1999;14:281.

9. Marrero JA, Su GL, Wei W, Emick D, Conjeevaram HS, Fontana RJ, et al.

Des‐gamma carboxyprothrombin can differentiate hepatocellular carcinoma

from nonmalignant chronic liver disease in american patients. Hepatology

2003;37:1114-21.

27

10. Kim DY, Paik YH, Ahn SH, Youn YJ, Choi JW, Kim JK, et al. PIVKA-II is a

useful tumor marker for recurrent hepatocellular carcinoma after surgical

resection. Oncology 2007;72:52-7.

11. Sato Y, Nakata K, Kato Y, Shima M, Ishii N, Koji T, et al. Early recognition

of hepatocellular carcinoma based on altered profiles of alpha-fetoprotein.

New England Journal of Medicine 1993;328:1802-6.

12. Shiraki K, Takase K, Tameda Y, Hamada M, Kosaka Y, Nakano T. A clinical

study of lectin‐reactive alpha‐fetoprotein as an early indicator of

hepatocellular carcinoma in the follow‐up of cirrhotic patients. Hepatology

2005;22:802-7.

13. Yamashita F, Tanaka M, Satomura S, Tanikawa K. Monitoring of

lectinreactive alpha-fetoproteins in patients with hepatocellular carcinoma

treated using transcatheter arterial embolization. Eur J Gastroenterol Hepatol

1995;7:627-33.

14. Yamashita F, Tanaka M, Satomura S, Tanikawa K. Prognostic significance of

Lens culinaris agglutinin A-reactive alpha-fetoprotein in small hepatocellular

carcinomas. Gastroenterology 1996;111:996.

15. Kumada T, Nakano S, Takeda I, Kiriyama S, Sone Y, Hayashi K, et al.

Clinical utility of Lens culinaris agglutinin-reactive alpha-fetoprotein in small

hepatocellular carcinoma: special reference to imaging diagnosis. Journal of

hepatology 1999;30:125.

28

16. Kusaba T. Relationship between Lens culinaris agglutinin reactive alpha-

fetoprotein and biological features of hepatocellular carcinoma. Kurume Med

J 1998;45:113-20.

17. Yamagata Y, Katoh H, Nakamura K, Tanaka T, Satomura S, Matsuura S.

Determination of α-fetoprotein concentration based on liquid-phase binding

assay using anion exchange chromatography and sulfated peptide introduced

antibody. Journal of immunological methods 1998;212:161-8.

18. Katoh H, Nakamura K, Tanaka T, Satomura S, Matsuura S. Automatic and

simultaneous analysis of Lens culinaris agglutinin-reactive α-fetoprotein ratio

and total α-fetoprotein concentration. Analytical chemistry 1998;70:2110-4.

19. Kagebayashi C, Yamaguchi I, Akinaga A, Kitano H, Yokoyama K, Satomura

M, et al. Automated immunoassay system for AFP–L3% using on-chip

electrokinetic reaction and separation by affinity electrophoresis. Analytical

biochemistry 2009;388:306-11.

20. Toyoda H, Kumada T, Tada T, Kaneoka Y, Maeda A, Kanke F, et al. Clinical

utility of highly sensitive Lens culinaris agglutinin‐reactive alpha‐fetoprotein

in hepatocellular carcinoma patients with alpha‐fetoprotein< 20 ng/mL.

Cancer science 2011;102:1025-31.

21. Bruix J, Sherman M. Management of hepatocellular carcinoma: an update.

Hepatology 2011;53:1020-2.

29

22. Representatives EGB. EASL–EORTC Clinical Practice Guidelines:

Management of hepatocellular carcinoma. Journal of hepatology

2012;56:908-43.

23. Makuuchi M, Kokudo N, Arii S, Futagawa S, Kaneko S, Kawasaki S, et al.

Development of evidence‐based clinical guidelines for the diagnosis and

treatment of hepatocellular carcinoma in Japan. Hepatology Research

2008;38:37-51.

24. Shinagawa T, Ohto M, Kimura K, Tsunetomi S, Morita M, Saisho H, et al.

Diagnosis and clinical features of small hepatocellular carcinoma with

emphasis on the utility of real-time ultrasonography. Gastroenterology

1984;86:1l.

25. Ikeda K, Saitoh S, Koida I, Tsubota A, Arase Y, Chayama K, et al. Diagnosis

and follow‐up of small hepatocellular carcinoma with selective intraarterial

digital subtraction angiography. Hepatology 1993;17:1003-7.

26. Takayasu K, Moriyama N, Muramatsu Y, Makuuchi M, Hasegawa H,

Okazaki N, et al. The diagnosis of small hepatocellular carcinomas: efficacy

of various imaging procedures in 100 patients. AJR. American journal of

roentgenology 1990;155:49-54.

27. Takayasu K, Furukawa H, Wakao F, Muramatsu Y, Abe H, Terauchi T, et al.

CT diagnosis of early hepatocellular carcinoma: sensitivity, findings, and CT-

pathologic correlation. AJR. American journal of roentgenology

1995;164:885-90.

30

28. Ebara M, Ohto M, Watanabe Y, Kimura K, Saisho H, Tsuchiya Y, et al.

Diagnosis of small hepatocellular carcinoma: correlation of MR imaging and

tumor histologic studies. Radiology 1986;159:371-7.

29. Bertino G, Ardiri A, Malaguarnera M, Malaguarnera G, Bertino N, Calvagno

GS. Hepatocellualar carcinoma serum markers. Seminars in oncology:

Elsevier; 2012. p.410-33.

30. Bolondi L, Sofia S, Siringo S, Gaiani S, Casali A, Zironi G, et al.

Surveillance programme of cirrhotic patients for early diagnosis and

treatment of hepatocellular carcinoma: a cost effectiveness analysis. Gut

2001;48:251-9.

31. Collier J, Sherman M. Screening for hepatocellular carcinoma. Hepatology

1998;27:273-8.

32. Trevisani F, D'Intino PE, Morselli-Labate AM, Mazzella G, Accogli E,

Caraceni P, et al. Serum α-fetoprotein for diagnosis of hepatocellular

carcinoma in patients with chronic liver disease: influence of HBsAg and

anti-HCV status. Journal of hepatology 2001;34:570-5.

33. Liebman HA, Furie BC, Tong MJ, Blanchard RA, Lo K-J, Lee S-D, et al.

Des-γ-carboxy (abnormal) prothrombin as a serum marker of primary

hepatocellular carcinoma. New England Journal of Medicine 1984;310:1427-

31.

34. Yuen M-F, Lai C-L. Serological markers of liver cancer. Best Practice &

Research Clinical Gastroenterology 2005;19:91-9.

31

35. Volk ML, Hernandez JC, Su GL, Lok AS, Marrero JA. Risk factors for

hepatocellular carcinoma may impair the performance of biomarkers: a

comparison of AFP, DCP, and AFP-L3. Cancer Biomarkers 2007;3:79-87.

36. Nakamura S, Nouso K, Sakaguchi K, Ito YM, Ohashi Y, Kobayashi Y, et al.

Sensitivity and specificity of des-gamma-carboxy prothrombin for diagnosis

of patients with hepatocellular carcinomas varies according to tumor size. The

American journal of gastroenterology 2006;101:2038-43.

37. Baek Y-H, Lee J-H, Jang J-S, Lee S-W, Han J-Y, Jeong J-S, et al. Diagnostic

role and correlation with staging systems of PIVKA-II compared with AFP.

Hepato-gastroenterology 2008;56:763-7.

38. França A, Elias Junior J, Lima B, Martinelli A, Carrilho F. Diagnosis, staging

and treatment of hepatocellular carcinoma. Brazilian journal of medical and

biological research 2004;37:1689-705.

39. Grizzi F, Franceschini B, Hamrick C, Frezza EE, Cobos E, Chiriva-Internati

M. Usefulness of cancer-testis antigens as biomarkers for the diagnosis and

treatment of hepatocellular carcinoma. J Transl Med 2007;5.

40. Semela D, Dufour J-F. Angiogenesis and hepatocellular carcinoma. Journal

of hepatology 2004;41:864-80.

41. Oda K, Ido A, Tamai T, Matsushita M, Kumagai K, Mawatari S-i, et al.

Highly sensitive lens culinaris agglutinin-reactive α-fetoprotein is useful for

early detection of hepatocellular carcinoma in patients with chronic liver

disease. Oncology reports 2011;26:1227-33.

32

42. Li D, Mallory T, Satomura S. AFP-L3: a new generation of tumor marker for

hepatocellular carcinoma. Clinica chimica acta 2001;313:15-9.

33

ABSTRACT (IN KOREAN)

간세포암 조기진단을 위한 종양표지자의 임상적 유용성

<지도교수 김도영>

연세대학교 대학원 의학과

임태섭

간세포암의 조기진단은 그 예후를 결정하는데 매우 중요한 역할을 한다.

Alpha-fetoprotein (AFP), Protein induce by vitamin K absence-II

(PIVKA-II), and Lens culinaris agglutinin-reactive fraction of AFP

(AFP-L3)는 간세포암을 진단할 수 있는 표지자자로 연구되어 왔다. 하지만,

어떠한 종양표지자가 간세포암의 조기진단에 가장 좋은 도구인지는 아직

의견이 일치된 바가 없다. 따라서 본 연구에서는, 간세포암의

조기진단에서 종양표지자들의 임상적 유용성에 대하여 알아보고자 한다.

환자군은 2012 년 1 월부터 2013 년 2 월까지 총 425 명의 환자 (간경변

196 명, 간세포암 229 명)가 포함되었고 종양표지자로 AFP, PIVKA-II, AFP-

L3 를 측정하였다. 본 환자들에서 간경변환자 중 58.7%, 간세포암중

76.0%가 남성이었고 간경변과 간암 환자들의 평균나이는 각각 55.8 세와

60.0 세였다. 모든 종양표지자들은 통계학적으로 유의하게 간경변

환자들과 비교하여 간암환자들에서 증가되어 있었다 (p = <0.001). 간경변

환자들로부터 간암을 진단하는데 있어 Area under receiver operating

characteristic curves (AUROC)는 AFP, PIVKA-II, AFP-L3 가 각각 0.679

(95% confidence interval [CI], 0.626-0.732, p = <0.001), 0.812 (95%

34

CI, 0.770-0.854, p = <0.001), 0.690 (95% CI, 0.638-0.742, p =

<0.001)였다. 3cm 미만의 단일종양의 조기 간세포암을 진단하는데

있어서도 PIVKA-II (AUROC = 0.705, 95% CI, 0.621-0.789, p = <0.001)가

AFP (AUROC = 0.623, 95% CI, 0.527-0.719, p = <0.019) 나 AFP-L3 (AUROC

= 0.561, 95% CI, 0.453-0.668, p = 0.245)보다 우월한 것으로 나타났다.

하지만, 조기간암을 진단하는데 있어 PIVKA-II (cut-off 40mAU/ml)는

25.6%의 낮은 민감도를 보였는데 이것은 AFP 를 조합함으로서 48.7%의

향상된 민감도를 보였고, AFP (cut-off 20ng/ml), PIVKA-II (cut-off 40

mAU/ml), AFP-L3 (cut-off 10%)를 조합할 때에는 56.4%까지 민감도의

향상을 보였다. AFP 이 20ng/ml 미만인 환자들에서는 PIVKA-II 의 AUROC

(0.743, 95% CI, 0.678-0.807; p = <0.001) 가 AFP-L3 (0.576, 95% CI,

0.500-0.653; p = 0.052)보다 우월한 것으로 나타났다. 또한, 본

연구에서는 로지스틱회귀분석을 통하여 AFP-L3 가 AFP 가 20ng/ml 이상인

AFP 위양성 간경화환자들로부터 간세포암환자를 구분하는데 유용한

지표임을 보여주었다. (odds ratio 1.076, 95% CI, 1.037-1.116, p = <

0.001). AFP, PIVKA, AFP-L3 등 모든 종양표지자들은 통계학적으로

유의하게 간세포암의 크기 및 병기와 연관되어 증가하는 것으로 나타났다.

결론적으로 AFP 와 PIVKA-II 를 조합하는 것이 조기간세포암을 진단하는데

유용한 도구로 사용될 수 있다고 생각되며, AFP-L3 는 AFP 위양성

환자들로부터 간세포암환자를 구분하는데 추가적인 역할을 할 수 있을

것으로 생각된다.

----------------------------------------------------------------------------------------

핵심되는 말: 간세포암, AFP, PIVKA-II, AFP-L3, 종양표지자

Clinical utility of tumor markers in early diagnosis of ...

Documents